Living on the edge: The dawn of edge computing
During my nearly three decades in the IT industry, I’ve helped organizations navigate changes to both the “what” and the “how” of computing. Now the “where” is changing.
In the days of yore, after I migrated from my abacus to the mainframe, and even later in the client/server era, most processing was centralized in data centers or hosting locations. Even with the advent of cloud, commercial or otherwise, computing was still mostly centralized, just located in someone else’s central data center.
Now, this is changing as well. Welcome to the edge.
What is edge computing?
In a nutshell, edge computing is the movement of applications and processing closer to the source of the data and out of largely centralized locations
Edge computing has application in things like industrial automation–factories with smart devices all streaming data on their health and productivity levels, and requiring some level of action. Or connected/autonomous vehicles, the safety of which depends on their ability to interact nearly instantaneously with other vehicles, smart traffic systems, and most importantly, people.
The Internet of Things (IoT), and the sheer volume of information being gathered from intelligent, connected devices, is one of the most obvious and common trends driving this shift. Sometimes, it’s as simple as aggregating data on the edge before sending it to the cloud or data center for processing. In other cases, IoT devices require quick, real-time responses that can’t be hampered by the latency involved with the round trip to a central location and back.
But low latency isn’t the only driver for edge. There are other less exotic motivations as well:
- Data protection: For various reasons, a company may not want to move its data to the cloud, but it wants to consume cloud resources.
- Data sovereignty: Processing information close to the source can keep data within borders and compliant with data sovereignty laws.
- Lack of a persistent connection: Airplanes, ships, and even fleets of cars can generate a phenomenal amount of data, and while these endpoints can connect via cell or satellite, the bandwidth to process this data just isn’t available today. Edge computing solves this problem. When a plane is on the ground, a ship’s in port, or a fleet’s in the garage, information can be gathered and consulted for real-time decisions, or it can be aggregated and sent to the big cloud in the data center for other purposes.
- Real-time decision making: A more relatable example is found in retail. I often use the example of the movie “Minority Report.” As the main character walked through the mall, biometric scanners identified him and made targeted marketing decisions in real time. Today’s technology makes fiction reality. While biometric identification is more often a use case for police, cell phones and wearables enable the same kind of real-time decisions in commercial settings.
The use cases for edge computing are virtually limitless.
What does edge computing look like?
It can be as simple as a server in a closet, but all of the major cloud providers have recognized the emerging trend and now provide offerings that allow their customers to consume some of the same services through edge deployments. AWS has Greengrass, Microsoft offers Azure Stack, and Google will offer Hybrid Cloud System.
Some of these are very prescriptive in how they are deployed, but essentially they’re providing a subset of their commercial cloud services for consumption on-prem. Your edge deployment can also be a “roll your own” implementation using whatever hardware, software, and services your implementation requires.
If your applications are running on VMware or OpenStack, these are perfectly viable edge solutions. VMware, in partnership with AWS, has announced an OVA template that includes Greengrass.
The takeaway is your edge deployment doesn’t have to fit into a prescribed model – these exist to simplify edge deployments. Your edge deployment needs to meet your requirements.
What does the future hold for edge computing?
Last year, Gartner blogged that “The Edge Will Eat the Cloud,” just as the cloud is eating traditional data center computing. Many companies agree, as evidenced by the cloud vendors that are providing offerings for the edge.
OEMs are investing in the cloud too. Dell, HP, and Lenovo are partnering to provide solutions on edge, and some, like Microsoft and Dell, are actively investing in IoT/edge development.
Given my tenure in the telecommunications industry, I would be remiss if I didn’t mention how network advancements are enabling edge as well. 5G will help bring edge about by providing high-bandwidth, low-latency capabilities to the IoT world over the air. No wires required, and it’s just around the corner!
Do you have an edge use case or questions about how the cloud or edge computing can assist your business? SHI’s Cloud and Innovative Solutions team can help. Contact your SHI account team to learn how.